babylm-baseline-100m-gpt-bert-mixed / tokenizer_config.json
lgcharpe's picture
Upload fully trained model (1B tokens)
462ad22 verified
raw
history blame
230 Bytes
{
"tokenizer_class": "PreTrainedTokenizerFast",
"bos_token": "<s>",
"eos_token": "</s>",
"unk_token": "<unk>",
"sep_token": "</s>",
"pad_token": "<pad>",
"cls_token": "<s>",
"mask_token": "<mask>"
}